Chris Sutton vs AI Tips: Premier League Predictions 2026
The smartest way to use Premier League predictions this weekend is to treat every scoreline as a price, not a promise. BBC Sport's Chris Sutton is forecasting all 380 top-flight games against AI, read...
Chris Sutton vs AI Tips: Premier League Predictions 2026
The smartest way to use Premier League predictions this weekend is to treat every scoreline as a price, not a promise. BBC Sport's Chris Sutton is forecasting all 380 top-flight games against AI, readers and guests, and week four pits him against Manchester City fan and guitarist Johnny Marr, with Manchester United's Fulham game on Sunday the headline fixture before a three-week international break. BBC scoring pays 10 points for the right result and 40 for the exact score, so a bold scoreline only beats a safe one if you land it often enough. Meanwhile footballpredictions.com has pencilled in Arsenal 2-0 Leeds United and Aston Villa 1-1 Brentford for 10 October 2026. Compare two or three sources, note where they disagree, and only back the picks where your own number differs from the market. Pitch Notes tracks these calls daily, so check back before kickoff.
Ever opened five tabs on a Friday night, found five different scorelines for the same match, and then felt oddly worse about your own pick? Sorry, that's basically my whole routine, and I suspect it's yours too, partner. The problem usually isn't a shortage of predictions. Columnists, models, crowd polls and tipster pages are everywhere, and almost none of them tell you how much to trust them. I track turnover, rebates and net position in a spreadsheet far more carefully than is healthy (please don't judge), and the lesson I keep relearning is that a prediction only becomes useful once you can price it. So in this guide from Pitch Notes, I'll walk through how weekend predictions used to work, what changed in 2026, how the scoring rules should reshape your picks, and what the next few fixtures actually look like, including Manchester United against Fulham and the 10 October games. I'll hedge where I'm unsure, because honestly nobody knows what Aston Villa will do on a wet night. Ready? Let's go plotting together.
Before 2025: how did Premier League weekend predictions work?
Before 2025, weekend Premier League predictions mostly meant one columnist's scoreline in a newspaper or on a TV panel, published once, rarely tracked, and never scored against anyone else. Readers had opinions and no scoreboard, so a confident pick and a lucky pick looked identical until Monday morning.
Looking back, the old routine was simple, maybe too simple. You read a columnist, glanced at a panel, and carried one scoreline into the weekend. If the pick missed, nobody really kept score, and if it hit, it got remembered forever. That asymmetry is the quiet flaw I'd flag, and sorry if it sounds harsh: without a record, reputation grows from highlights. In my own notes from that era, the only numbers I could verify were my own stakes and returns, never the source's hit rate. Odds comparison sites existed, sure, but the opinion layer sat apart from the price layer, so a tip like "City to win comfortably" came with no way to check whether the market had already priced it in. The practical result was that readers copied confidence rather than evidence. Does that sound familiar? I think many of us treated a famous name as a substitute for a track record, and a ten-match Premier League round only made the habit stronger, because there was always another game to move on to before the last pick was judged. For more on building a basic record, see our [Internal Link: beginner's guide to tracking football predictions].
The 2026 shift: what actually changed?
The big shift is accountability. BBC Sport now has Chris Sutton predicting all 380 Premier League games against AI, readers and rotating guests, with a public scoreboard, while prediction sites publish a scoreline early and flesh out the analysis later. Predictions became a scored contest instead of a one-off opinion.
According to BBC Sport, week four of the season sees Sutton take on Johnny Marr, whose fifth solo album, The Age Of Everything, is out on 2 October. The most popular scoreline picked by readers feeds the scoreboards and tables, which means the crowd now has a public, scored consensus. The scoring is blunt, and BBC Sport spells it out: "A correct result (picking a win, draw or defeat) is worth 10 points. The exact score earns 40 points." Prediction sites went in a different direction. footballpredictions.com lists Arsenal against Leeds United at 2-0 and Aston Villa against Brentford at 1-1, both dated 10 October 2026, with the note that "the full analysis will be added closer to kickoff." Here's the first thing I'd tell a friend: that early scoreline is a placeholder, not a conclusion, so a pick published days out can't have absorbed team news yet. Also, the listed 19:30 and 22:00 slots don't look like standard UK kickoffs, so I'd confirm times on the Premier League fixture list before planning around them.
Here is what the shift looks like in practice:
Scored contests: Sutton versus AI, readers and guests across all 380 games, with points on the table.
Crowd consensus: the most popular reader scoreline feeds the public scoreboards.
Early picks, late analysis: tipster pages post a scoreline first and add the reasoning near kickoff.
Side games: BBC Sport's Match of the Day predictor adds six fixtures a week from domestic and European competitions.
What changed for players?
For players, the scoring rule matters more than the pundit. BBC Sport pays 10 points for a correct result and 40 for the exact score, so exact scorelines are worth four times a plain result. That reward structure punishes lazy draw picks and rewards modest, likely scorelines such as 1-0 and 2-1.
Let me show the maths, apologies in advance for the spreadsheet energy. Assume, purely as an illustration and not as measured data, that a home side has a 40 percent chance to win and that 1-0 lands 11 percent of the time. Under the BBC rules, a 1-0 pick returns about 7.3 expected points (0.11 times 40, plus 0.29 times 10). Now suppose a draw has a 26 percent chance and 1-1 makes up 44 percent of all draws. A 1-1 pick returns only about 6.0. That means a draw call needs roughly a 31 percent draw probability just to match the home-win call. This is the part most tipster pages never mention: the scoring rule, not the pundit, decides which kind of pick is cheap. The same logic applies if real money is involved. A correct-score market pays more but hits far less often, and in my experience your turnover grows much faster than your net position if you keep chasing it. If you want the deeper version, I wrote it up in [Internal Link: expected value for correct-score picks].
My routine, for what it's worth, is this:
- Write your own win, draw and loss percentages before reading anyone else's pick.
- Compare at least two sources and note where they disagree, since disagreement is where the price is soft.
- Within your chosen result, choose the most common scoreline, usually 1-0, 2-1 or 2-0, instead of a headline-friendly 4-2.
- Skip any game where you disagree with every source but can't say why.
What does this mean now for this weekend's fixtures?
This weekend, the pressure fixture is Manchester United against Fulham on Sunday, with Michael Carrick under scrutiny before a three-week international break. Manchester City sit on a 100 percent start under Enzo Maresca. Treat both as narrative-heavy games, where the story can distort the price more than the football does.
Take Sunday first. BBC Sport reports that pressure is growing on Carrick, and Sutton's view, paraphrased, is that teams with a bad start can see negativity fester over the break, with speculation about the manager's future getting louder. That makes the fixture emotionally loud, but loud isn't the same as predictable. A cautious reading: when a story is this dramatic, public scorelines tend to lean toward a dramatic result, so the unglamorous low-scoring outcome can be the softer spot. My contrarian note, and I could be wrong, is that the loudest fixture of the weekend is rarely the best pick. Next, City's perfect record includes the win at Old Trafford, where Marr says Phil Foden was sent off early and the team still dug in. Perfect starts inflate crowd scorelines too, because everyone piles onto the winners. Finally, on 10 October, the early lines from footballpredictions.com are Arsenal 2-0 Leeds United and Aston Villa 1-1 Brentford. By my numbers, Arsenal 2-0 is a sensible modal scoreline, while Villa 1-1 Brentford is exactly the draw call that needs about a 31 percent draw probability to justify itself, so I'd probably pass, sorry. Our [Internal Link: weekly Premier League fixture previews] cover team news as it lands.
One more point from the investor side of my brain. Predictions are free, but acting on them isn't, so please keep your stakes small, set your limits before kickoff, and never chase a Saturday loss on Sunday. The BeGambleAware site has practical tools if you ever feel the habit slipping, and nothing here is a guarantee of any result. I log every pick as a rough percentage, then compare it with the actual outcome at the end of each month, and the main thing I've learned is that my net position moves far more with discipline than with any single scoreline. Treat a tipster scoreline as an input to your own number, not a replacement for it. If a source can't show you a track record, I'd weight it lightly, and if it does show one, I'd still ask how many picks the sample covers, because forty games is a tiny sample and 380 is a much more honest one. Adults only, 18+, and play for fun first.
Three predictions for next quarter
Here are my three guesses for the rest of 2026, offered with the usual caveat that football loves to embarrass anyone who sounds too sure.
- Scored scoreboards spread. If Sutton versus AI, readers and guests keeps drawing attention, I expect more outlets to publish running records instead of one-off picks, which would finally make "who is actually good at this?" a checkable question.
- Late analysis beats early scorelines. Pages that post a number days ahead and add reasoning near kickoff will keep losing ground to sources that update after team news, so the freshest pick should usually outrank the earliest one.
- Exact-score hit rates stay low. With 40 points on offer, the crowd's favourite scoreline will keep being 1-0 or 2-1 in most games, and players who chase 3-2 will mostly lose to players who stay boring.
The wider context helps too. Pitch Notes grew out of our FIFA World Cup coverage, and Los Angeles alone hosted eight matches and 39 days of fan celebrations, which tells me how much appetite there is for football content that explains itself. That appetite doesn't disappear when the tournament ends. It just moves back to the league. So my plan for this weekend is simple: build my own probabilities, compare them with BBC Sport and the tipster pages, take only the picks where my number differs, and write everything down. If you do the same, you'll at least know why you won or lost, which is more than most of us could say in the old newspaper era. See you at kickoff, partner.
Frequently Asked Questions
Q: What is the best way to use Premier League predictions this weekend?
A: Use them as inputs to your own probabilities, not as final answers. Write your own win, draw and loss percentages first, then compare two or three sources such as BBC Sport and footballpredictions.com. Back a pick only where your number differs from the crowd, and skip games where you can't explain the disagreement. That keeps decisions consistent and makes your results easy to review afterwards.
Q: How does BBC Sport score its predictions?
A: BBC Sport awards 10 points for a correct result and 40 points for the exact score. A correct result means picking the win, draw or defeat. The most popular scoreline chosen by readers is used in the scoreboards and tables. Because the exact score is worth four times a plain result, sensible scorelines like 1-0 and 2-1 tend to make the most sense.
Q: Are early predictions from tipster sites reliable?
A: They are best treated as placeholders until the analysis arrives. footballpredictions.com, for example, states that full analysis is added closer to kickoff, so a scoreline posted days ahead cannot include late team news. Check the page again on matchday, confirm kickoff times against the official Premier League fixture list, and weight the later update more heavily.
Q: Is a 1-1 draw pick ever worth it?
A: A 1-1 pick is worth it only when you believe a draw is genuinely likely. Under a 10 and 40 scoring system, my illustrative maths suggests a draw call needs roughly a 31 percent draw probability to match a 40 percent home-win call on 1-0. Without that edge, a modest home or away scoreline usually returns more expected points over a long season.
Q: Why do my picks keep missing the exact score?
A: Exact scores are rare events, so frequent misses are normal. Even a well-judged 1-0 or 2-0 pick may only land around one game in nine or ten, and bolder scorelines land less often. Keep a monthly log of both result accuracy and exact-score accuracy, and judge yourself on result accuracy first, because that is where skill shows up fastest.
Q: How much should I stake on correct-score picks?
A: Keep stakes very small, because correct-score outcomes hit far less often than simple win, draw or loss results. A common rule of thumb is a small fraction of your bankroll per pick, set before the weekend begins, and never increased to recover a loss. Betting is for adults aged 18 and over, and support is available from BeGambleAware if it stops being fun.
Thank you for reading.
Pitch Notes · Editorial Archive · 2026